The Four Business Models We Build Software For

Four shapes, four different centres of gravity

After enough projects, the briefs start to rhyme. Almost everything we build falls into one of four business shapes, and each has a characteristic place where the value is — and a characteristic mistake.

1. Service businesses: capacity is the product

Agencies, consultancies, clinics, trades, professional services. What you sell is people's time, so the constraining resource is capacity and the software that matters is whatever makes capacity visible before it is committed.

The mistake: building a beautiful CRM to win more work while having no reliable view of whether the work can be delivered. Selling capacity you do not have is more damaging than not selling it at all.

• Scheduling that knows about skills, not just availability
• Utilisation that is visible before quoting, not reported after
• Time capture that takes seconds, or it will not happen
• Project profitability while the project is live, not at the end

2. Product businesses: inventory truth is everything

Wholesale, retail, manufacturing, distribution. Everything hangs off one number being right: what you actually have. Get that wrong and every downstream system amplifies the error into oversells, credit notes and apology emails.

The mistake: adding sales channels before the stock figure is trustworthy. Each new channel multiplies an existing inaccuracy rather than adding revenue cleanly.

One wholesale client was writing off two to three thousand pounds a week in credits traceable to stock errors. The fix was not a new sales channel — it was one honest number.

3. Marketplaces: trust before liquidity

Two-sided platforms are the hardest of the four, because the software is only half the problem. The other half is the cold-start problem, and no amount of engineering solves it.

The mistake: building the full platform before proving a single transaction can be matched manually. The marketplaces that work usually ran the first fifty matches by hand in a spreadsheet, then automated what they learned.

• Identity and reputation, because trust is the actual product
• Payments with escrow or staged release
• Dispute handling — the feature everyone defers and everyone needs
• Matching, which is where a marketplace earns its take rate

4. Internal operations: the re-keying tax

Any business where the software serves staff rather than customers. The value here is almost always the same: eliminate the places where a person moves data between systems by hand.

The mistake: building a comprehensive internal platform when three targeted integrations would deliver most of the benefit for a fifth of the cost. Internal tools have a strong tendency to sprawl, because everyone who sees a demo has one more idea.

How the shapes change the budget

Internal operations is consistently the cheapest and the fastest to pay back, which is why we suggest it first even to businesses whose ambitions lie elsewhere. A funded, working internal system builds the credibility to do the bigger thing.

Businesses that are two shapes at once

Plenty are. A manufacturer with a direct-to-consumer arm is a product business and an internal operations business simultaneously. A clinic with an online booking portal is a service business with a customer-facing surface.

In those cases we build the shared data layer once and treat each surface separately. What we avoid is a single system trying to be both, which reliably produces software that is mediocre at each.

Frequently asked questions